CN103472031A - Navel orange sugar degree detection method based on hyper-spectral imaging technology - Google Patents
Navel orange sugar degree detection method based on hyper-spectral imaging technology Download PDFInfo
- Publication number
- CN103472031A CN103472031A CN2013104289251A CN201310428925A CN103472031A CN 103472031 A CN103472031 A CN 103472031A CN 2013104289251 A CN2013104289251 A CN 2013104289251A CN 201310428925 A CN201310428925 A CN 201310428925A CN 103472031 A CN103472031 A CN 103472031A
- Authority
- CN
- China
- Prior art keywords
- navel orange
- detection method
- pol
- sugar degree
- method based
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Landscapes
- Investigating Or Analysing Materials By Optical Means (AREA)
Abstract
The invention relates to a navel orange sugar degree detection method based on a hyper-spectral imaging technology. The navel orange sugar degree detection method based on the hyper-spectral imaging technology comprises the following steps: collecting spectra of a navel orange through a hyper-spectral imaging system, and performing black-white calibration to eliminate the noise influence; measuring the sugar degree of the navel orange through a digital refractometer; establishing a relationship between the spectra and the sugar degree so as to facilitate subsequent measurement of the sugar degree and nondestructive detection of the sugar degree of the navel orange. Compared with a conventional detection method, the navel orange sugar degree detection method based on the hyper-spectral imaging technology is high in detection speed, easy and convenient to operate and harmless to fruits. Compared with a near infrared spectrum detection method, the navel orange sugar degree detection method based on the hyper-spectral imaging technology has the advantages that more information is obtained and a detection result is more accurate and stable.
Description
Technical field
The present invention designs the detection method of fruit pol, a kind of navel orange pol detection method based on the high light spectrum image-forming technology of specific design.
Background technology
The judgement of fruit quality and detection are one of important topics of agricultural product research always.Along with the raising of people's quality of life, the consumer, when choosing fruit, except focusing on the external sorts such as size, color, face shaping, also very pays close attention to as mouthfeel, pol, acidity and vitamin content etc. for inside quality.The sweet taste of fruit is relevant with the height of its pol, and pol, as one of key factor that determines fruit internal quality, is paid close attention to by people always.Therefore, studying the method that quick, real-time not damaged detects pol in fruit is the important prerequisite of fruit grading.
Traditional pol detection method is directly measured fruit fruit juice sampling by the hand-held saccharimeter, have time-consuming, efficiency is low, destructive, easily produce the shortcoming such as operate miss.And there are some shortcomings in the near infrared spectrum diagnostic techniques, for example: need representative and sample that chemical score is known in a large number to set up model; The modeling capital is high, and the test expenditure is large.Dimension, the detection method of a kind of convenient, general fruit pol of exigence.The high light spectrum image-forming technology has the characteristics of multiband, high resolving power and collection of illustrative plates unification, can combine two dimensional image and spectral technique together, can overcome the some shortcomings of other detection methods, has more advantage.In recent years, the high light spectrum image-forming technology has obtained application widely in the fruit inside and outside quality detects.Therefore, it is feasible adopting high spectral technique to detect the fruit pol.
Summary of the invention
Purpose of the present invention: the deficiency existed for present fruit pol method of testing, the purpose of this invention is to provide a kind of navel orange pol detection method based on the high light spectrum image-forming technology, by the high-spectral data analytical technology, the spectrum collected is carried out to the extraction of characteristic spectrum parameter, set up correlation model, thereby the pol of fruit is carried out to the quantification prediction.
Technical scheme: in order to realize the foregoing invention purpose, the present invention take navel orange as the technology bill that example adopts as follows: a kind of fruit pol detection method based on high spectral technique comprises the following steps:
1) collection of high spectrum image and correction: utilize Hyperspectral imager to carry out spectra collection to navel orange, and carry out the black and white demarcation;
2) mensuration of pol: adopt digital refractometer to measure the navel orange pol;
3) choosing of image: the high spectrum image collected is selected to area-of-interest;
4) extraction of characteristic wavelength: utilize genetic algorithm (GA) to carry out smothing filtering to the averaged spectrum collected, and pick out suitable characteristic variable;
5) set up model: utilize partial least square method to carry out modeling to the select characteristic variable of GA algorithm, and utilize the checking sample to test;
6) utilize above-mentioned model to detect the fruit pol.
The collection of wherein said high spectrum image and correction, refer to utilize Hyperspectral imager to carry out spectra collection to navel orange, and carry out the black and white demarcation, eliminates noise effect.
The mensuration of wherein said pol, refer to adopt digital refractometer to measure the navel orange pol, and selected part is as the modeling collection from laboratory sample to utilize the K-S method, and the residue sample is as being the checking collection.
Choosing of wherein said image, refer to that the high spectrum image to collecting is selected area-of-interest, obtains the curve of spectrum of sample.
The extraction of wherein said characteristic wavelength refers to and utilizes genetic algorithm (GA) to carry out smothing filtering to the averaged spectrum collected, and picks out suitable characteristic variable.
The wherein said model of setting up, refer to utilize partial least square method to carry out modeling to the select characteristic variable of GA algorithm, and utilize the checking sample to test.
Utilize partial least square method to carry out modeling to the select characteristic variable of GA algorithm, the predicted value of model and the related coefficient of measured value and root-mean-square deviation are respectively 0.83 and 0.54.
Set up the relation between spectrum and pol, thus the measurement of pol after convenient, can be for the Non-Destructive Testing of navel orange pol.With the conventional sense method, compare, detection speed is fast, simple to operation, harmless to fruit; Detect and compare with near infrared spectrum, more comprehensively, testing result is more accurately with stable for the information obtained.
The invention has the beneficial effects as follows: the variation of the spectral signature that the present invention utilizes the navel orange sugar to cause, adopt partial least square method to detect the navel orange pol, a kind of detection method of fast and stable can be provided.
The accompanying drawing explanation
Fig. 1 process flow diagram of the present invention;
Fig. 2 Hyperspectral imager.
Wherein: 1. computing machine; 2. grating spectrograph; 3. CCD; 4. light box; 5. light source; 6. imaging lens; 7. transfer table; 8. sample; 9. bracing frame.
Fig. 3 is choosing of navel orange sample ROI.
The averaged spectrum reflected value curve that Fig. 4 is sample.
The result of Fig. 5 for utilizing genetic algorithm (GA) to be selected spectrum.
Fig. 6 is gained equation coefficient distribution plan when utilizing the partial least square method modeling.
Fig. 7 is forecast model training result figure.
Embodiment
Below in conjunction with drawings and Examples, the present invention is described further
Experiment material is the navel orange sample of buying from certain fruit wholesale market, by surface dirt wiped clean, number consecutively.Navel orange is placed under the experimental situation of 20 ℃ of temperature, humidity 60%, makes sample reach room temperature, disease is prevented and treated at least 24h.
The detection of navel orange pol:
high spectroscopic system and image acquisition: the image acquisition of carrying out of utilizing Fig. 2, in Fig. 2, structure comprises computing machine 1, grating spectrograph 2, CCD3, light box 4, light source 5, imaging lens 6, transfer table 7, sample 8, bracing frame 9, the upper end of light box 4 is connected with CCD3, CCD3 connects grating spectrograph 2, grating spectrograph 2 connects imaging lens 6, the two side of light box 4 is connected with respectively light source 5, the bottom of described light box 4 connects bracing frame 9, be placed with transfer table 7 on bracing frame 9, be placed with sample 8 on transfer table 7, described transfer table 7 is connected respectively computing machine 1 with CCD3.Experimental data for fear of baseline wander, is opened preheating 30min by Hyperspectral imager before gathering.Navel orange is placed on the black floor of a mobile platform, allows carpopodium keep horizontal direction.For making to avoid image fault, illumination is saturated.The time shutter that the Hyperspectral imager camera is set is 5 ms, and resolution ratio of camera head is 1344
1024, the translational speed of electric platforms is 16.6 mm/s, and spectral range is 400 ~ 1000 nm, spectral resolution is 2.8 nm, spectrum sample is spaced apart 2.44nm, collects 254 images under wavelength, finally obtains the high spectrum image data block that a size is 1344 * 1024 * 254.The high spectrum image collected is carried out to the black and white demarcation, eliminate noise effect.Scanning standard white correction plate obtains complete white uncalibrated image W, closes camera shutter and collects complete black uncalibrated image B, according to publicity (1), makes the absolute image I collected become relative image R.
obtaining of the curve of spectrum: for the high spectrum image collected, select the square shape area-of-interest Region of Interesting (ROI) of 7 * 7 at the navel orange center, as Fig. 3, then calculate the averaged spectrum of 49 pixels, obtain the curve of spectrum of sample, as Fig. 4.
the detection of navel orange pol: the content of navel orange sample pol adopts digital refractometer PR-101 α (Atago Co Ltd, Tokyo, Japan) to measure.Corresponding navel orange high spectrum image center is cut one, extrudes a fruit juice and drips on the test window of digital refractometer, and by the K-S method, from laboratory sample, selected part is as the modeling collection, and residue collects as checking.
characteristic wavelength is chosen: genetic algorithm (Genetic Algorithm) is at first to propose professor J.Holland by the U.S. in 1975, that the evolution laws (survival of the fittest, survival of the fittest genetic mechanism) that a class is used for reference organic sphere develops and next randomization searching method.Principal feature is directly operated structure objects, does not have the restriction of differentiate and continuous; There is inherent Implicit Parallelism and better global optimizing ability; Adopt the optimization method of randomization, the search volume that energy automatic acquisition and guidance are optimized, adjust the direction of search adaptively, do not need the rule of determining.In spectral analysis, can utilize genetic algorithm (GA) to realize global search, thereby eliminate the interference of irrelevant variable, select characteristic wavelength.
The averaged spectrum of ROI is imported to matlab R2010a, and parameter is set to: number of groups is 30, crossover probability 0.50, and compiling probability 0.01, iterations is 100, independent operating 100 times.According to above-mentioned parameter operation GA100 time, export 0-1 binary-coded character string at every turn, calculate the probability that wavelength points is designated " 1 ".
Pick out 46 wavelength variable point: 17-19 that frequency is higher, 33-35,38,40-42,51-56,82,84-92,94-96,
121-125,178-181,232-234,239-241,253-254。As Fig. 5.
modeling result: partial least square method is a kind of mathematical optimization technology, and it finds the optimal function coupling of one group of data by the quadratic sum of minimum error.The select characteristic variable of GA algorithm is carried out to modeling.Wherein, equation coefficient distributes as Fig. 6, utilizes form to enumerate out coefficient, as ascending as 1(46 wavelength of table).The equation intercept is 10.570526 again, so obtain utilizing the resulting equation of partial least square method:
In formula:
be i the corresponding spectral reflectivity of wavelength points;
be i the corresponding equation coefficient of wavelength points; Z is pol.
Table 1 gained model equation coefficient
Utilize the reliability of residue Sample model, training result as shown in Figure 7.Visible after smoothing processing, the related coefficient of the forecast set that the GA-PLS model obtains and root-mean-square deviation are respectively 0.83 and 0.54.Visible this model that can utilize is predicted the navel orange pol.
Claims (3)
1. the navel orange pol detection method based on the high light spectrum image-forming technology is characterized in that comprising the following steps:
1) collection of high spectrum image and correction: utilize Hyperspectral imager to carry out spectra collection to navel orange, and carry out the black and white demarcation, eliminate noise effect;
2) mensuration of pol: adopt digital refractometer to measure the navel orange pol;
3) choosing of image: the high spectrum image collected is selected to area-of-interest, obtain its curve of spectrum;
4) extraction of characteristic wavelength: utilize genetic algorithm (GA) to carry out smothing filtering to the averaged spectrum collected, and pick out suitable characteristic variable;
5) set up model: utilize partial least square method to carry out modeling to the select characteristic variable of GA algorithm, and utilize the checking sample to test;
6) utilize above-mentioned model to detect the fruit pol.
2. a kind of navel orange pol detection method based on the high light spectrum image-forming technology according to claim 1, it is characterized in that Hyperspectral imager comprises computing machine (1), grating spectrograph (2), CCD(3), light box (4), light source (5), imaging lens (6), transfer table (7), sample (8), bracing frame (9), the upper end of light box (4) is connected with CCD(3), CCD(3) connect grating spectrograph (2), grating spectrograph (2) connects imaging lens (6), the two side of light box (4) is connected with respectively light source (5), the bottom of described light box (4) connects bracing frame (9), be placed with transfer table (7) on bracing frame (9), be placed with sample (8) on transfer table (7), described transfer table (7) and CCD(3) be connected respectively computing machine (1).
3. according to claim 1,2 described a kind of navel orange pol detection methods based on the high light spectrum image-forming technology, the time shutter that it is characterized in that arranging the Hyperspectral imager camera is 5 ms, and resolution ratio of camera head is 1344
1024, the translational speed of electric platforms is 16.6 mm/s, and spectral range is 400 ~ 1000 nm, spectral resolution is 2.8 nm, spectrum sample is spaced apart 2.44nm, collects 254 images under wavelength, finally obtains the high spectrum image data block that a size is 1344 * 1024 * 254.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN2013104289251A CN103472031A (en) | 2013-09-20 | 2013-09-20 | Navel orange sugar degree detection method based on hyper-spectral imaging technology |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN2013104289251A CN103472031A (en) | 2013-09-20 | 2013-09-20 | Navel orange sugar degree detection method based on hyper-spectral imaging technology |
Publications (1)
Publication Number | Publication Date |
---|---|
CN103472031A true CN103472031A (en) | 2013-12-25 |
Family
ID=49796971
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN2013104289251A Pending CN103472031A (en) | 2013-09-20 | 2013-09-20 | Navel orange sugar degree detection method based on hyper-spectral imaging technology |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN103472031A (en) |
Cited By (12)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103822711A (en) * | 2014-03-03 | 2014-05-28 | 中国科学院遥感与数字地球研究所 | Digital image display method and hyper-spectral telescope |
CN105158178A (en) * | 2015-10-08 | 2015-12-16 | 华中农业大学 | Rapid modeling method for detecting sugar content of navel orange based on spectral peak area in high spectral transmission technology |
CN105181606A (en) * | 2015-08-28 | 2015-12-23 | 中国农业科学院农产品加工研究所 | Hyperspectral imaging technology-based method for detecting sucrose content distribution of peanut |
CN105181611A (en) * | 2015-10-08 | 2015-12-23 | 华中农业大学 | Nondestructive testing device for hyperspectral transmission imaging of sphere-like fruits |
CN105241824A (en) * | 2015-09-30 | 2016-01-13 | 江苏大学 | Method for quantitatively detecting solid fermentation index distribution difference through hyperspectral image technology |
CN105717051A (en) * | 2016-04-22 | 2016-06-29 | 合肥美菱股份有限公司 | System capable of rapidly detecting fruit and vegetable freshness and refrigerator |
CN106067173A (en) * | 2016-05-30 | 2016-11-02 | 湖南生物机电职业技术学院 | The Complexity Measurement lossless detection method of citrusfruit pol |
CN107607480A (en) * | 2016-07-12 | 2018-01-19 | 湖南生物机电职业技术学院 | The lossless detection method of navel orange effective acidity |
CN109827910A (en) * | 2019-01-22 | 2019-05-31 | 塔里木大学 | A kind of quick monitoring process method of orchard establishing data |
CN111982835A (en) * | 2020-08-17 | 2020-11-24 | 吉林求是光谱数据科技有限公司 | Fruit sugar degree nondestructive testing device and method based on silicon-based multispectral chip |
CN112964719A (en) * | 2021-04-26 | 2021-06-15 | 山东深蓝智谱数字科技有限公司 | Hyperspectrum-based food fructose detection method and device |
CN113554575A (en) * | 2020-04-23 | 2021-10-26 | 华东交通大学 | High-reflection object surface highlight removing method based on polarization principle |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH01161123A (en) * | 1987-12-16 | 1989-06-23 | Shimadzu Corp | Deciding method for similarity of spectrum pattern |
JP2001289783A (en) * | 2000-04-10 | 2001-10-19 | Ishikawajima Harima Heavy Ind Co Ltd | Method and device for measuring so3 concentration in exhaust gas |
CN1865930A (en) * | 2006-05-09 | 2006-11-22 | 江西农业大学 | Reflection type fast detector for acidity of fruit confection |
CN102495005A (en) * | 2011-11-17 | 2012-06-13 | 江苏大学 | Method for diagnosing crop water deficit through hyperspectral image technology |
-
2013
- 2013-09-20 CN CN2013104289251A patent/CN103472031A/en active Pending
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH01161123A (en) * | 1987-12-16 | 1989-06-23 | Shimadzu Corp | Deciding method for similarity of spectrum pattern |
JP2001289783A (en) * | 2000-04-10 | 2001-10-19 | Ishikawajima Harima Heavy Ind Co Ltd | Method and device for measuring so3 concentration in exhaust gas |
CN1865930A (en) * | 2006-05-09 | 2006-11-22 | 江西农业大学 | Reflection type fast detector for acidity of fruit confection |
CN102495005A (en) * | 2011-11-17 | 2012-06-13 | 江苏大学 | Method for diagnosing crop water deficit through hyperspectral image technology |
Non-Patent Citations (2)
Title |
---|
刘燕德等: "基于高光谱的GA和SPA算法对赣南脐橙叶绿素定量分析", 《光谱学与光谱分析》, vol. 32, no. 12, 31 December 2012 (2012-12-31) * |
郭恩有等: "脐橙糖度的高光谱图像无损检测技术", 《农业机械学报》, vol. 39, no. 5, 31 May 2008 (2008-05-31) * |
Cited By (14)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103822711B (en) * | 2014-03-03 | 2015-12-02 | 中国科学院遥感与数字地球研究所 | Digital image display methods and EO-1 hyperion telescope |
CN103822711A (en) * | 2014-03-03 | 2014-05-28 | 中国科学院遥感与数字地球研究所 | Digital image display method and hyper-spectral telescope |
CN105181606A (en) * | 2015-08-28 | 2015-12-23 | 中国农业科学院农产品加工研究所 | Hyperspectral imaging technology-based method for detecting sucrose content distribution of peanut |
CN105241824A (en) * | 2015-09-30 | 2016-01-13 | 江苏大学 | Method for quantitatively detecting solid fermentation index distribution difference through hyperspectral image technology |
CN105181611B (en) * | 2015-10-08 | 2018-02-23 | 华中农业大学 | Spherical fruit transmits high light spectrum image-forming the cannot-harm-detection device |
CN105181611A (en) * | 2015-10-08 | 2015-12-23 | 华中农业大学 | Nondestructive testing device for hyperspectral transmission imaging of sphere-like fruits |
CN105158178A (en) * | 2015-10-08 | 2015-12-16 | 华中农业大学 | Rapid modeling method for detecting sugar content of navel orange based on spectral peak area in high spectral transmission technology |
CN105717051A (en) * | 2016-04-22 | 2016-06-29 | 合肥美菱股份有限公司 | System capable of rapidly detecting fruit and vegetable freshness and refrigerator |
CN106067173A (en) * | 2016-05-30 | 2016-11-02 | 湖南生物机电职业技术学院 | The Complexity Measurement lossless detection method of citrusfruit pol |
CN107607480A (en) * | 2016-07-12 | 2018-01-19 | 湖南生物机电职业技术学院 | The lossless detection method of navel orange effective acidity |
CN109827910A (en) * | 2019-01-22 | 2019-05-31 | 塔里木大学 | A kind of quick monitoring process method of orchard establishing data |
CN113554575A (en) * | 2020-04-23 | 2021-10-26 | 华东交通大学 | High-reflection object surface highlight removing method based on polarization principle |
CN111982835A (en) * | 2020-08-17 | 2020-11-24 | 吉林求是光谱数据科技有限公司 | Fruit sugar degree nondestructive testing device and method based on silicon-based multispectral chip |
CN112964719A (en) * | 2021-04-26 | 2021-06-15 | 山东深蓝智谱数字科技有限公司 | Hyperspectrum-based food fructose detection method and device |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN103472031A (en) | Navel orange sugar degree detection method based on hyper-spectral imaging technology | |
Mishra et al. | Close range hyperspectral imaging of plants: A review | |
Ali et al. | Estimating leaf functional traits by inversion of PROSPECT: Assessing leaf dry matter content and specific leaf area in mixed mountainous forest | |
Qin et al. | Hyperspectral and multispectral imaging for evaluating food safety and quality | |
CN110763698B (en) | Hyperspectral citrus leaf disease identification method based on characteristic wavelength | |
CN103278460B (en) | A kind of mandarin tree red spider herbivore stress situation method for testing and analyzing | |
CN100357725C (en) | Method and device for rapidly detecting tenderness of beef utilizing near infrared technology | |
CN104697943A (en) | Rape water stress nondestructive testing method | |
CN108956604A (en) | A method of Eriocheir sinensis quality is identified based on hyper-spectral image technique | |
CN105203464A (en) | Method for detecting oleic acid content distribution in peanuts based on hyperspectral imaging technology | |
CN105021617A (en) | Hyperspectral imaging-based measuring equipment of chlorophyll content of whole rice plant and method thereof | |
CN109540836A (en) | Near infrared spectrum pol detection method and system based on BP artificial neural network | |
Xiong et al. | A micro-damage detection method of litchi fruit using hyperspectral imaging technology | |
CN105181606B (en) | The method that cane sugar content is distributed in peanut is detected based on high light spectrum image-forming technology | |
Feng et al. | A hyperspectral imaging system for an accurate prediction of the above-ground biomass of individual rice plants | |
CN111751376A (en) | Rice nitrogen nutrition estimation method based on canopy image feature derivation | |
Gomes et al. | Prediction of sugar content in port wine vintage grapes using machine learning and hyperspectral imaging | |
CN104677853A (en) | Fresco scaling damage assessment method based on near-infrared hyperspectrum | |
Mu et al. | Non‐destructive detection of blueberry skin pigments and intrinsic fruit qualities based on deep learning | |
CN107576600A (en) | A kind of quick determination method for smearing tea grain size category | |
Li et al. | Retrieval of nitrogen content in apple canopy based on unmanned aerial vehicle hyperspectral images using a modified correlation coefficient method | |
Lin et al. | Outdoor detection of the pollution degree of insulating materials based on hyperspectral model transfer | |
CN106556602A (en) | A kind of detection method and its device of fish body freshness | |
Zhou et al. | Hyperspectral imaging technology for detection of moisture content of tomato leaves | |
CN115184281B (en) | Method and system for determining concentration of solution components based on two-dimensional spectrum |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
C12 | Rejection of a patent application after its publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20131225 |